A Water Supplier's View
An integrated strategy for dealing with pesticide pollution of drinking water catchments

This paper describes an integrated five-point plan for cost effectively meeting the stringent EC Drinking Water Standard for pesticides of 0.1 µg/l, in the supply area of Severn Trent Water—a leading water supply company in the UK. The area is supplied from a mix of water resource types, and particular problems have been encountered with diffuse herbicide pollution of the river derived catchments. The strategies being implemented include pesticide use surveys; a comprehensive pesticide monitoring programme; promotion of a high profile catchment protection measures—the Spraysafe campaign; research on catchment vulnerability assessment using computer based geographic models; and water treatment to remove residual contaminants. By Bob Breach and Mike Porter.

Introduction

The European Commission (EC) standard for any pesticide in drinking water, irrespective of its toxicity, is 0.1 µg/l. This value, which represents one part in ten billion, is effectively a surrogate standard for zero. By contrast, health related guideline values for pesticides recently issued by the World Health Organisation are in nearly all cases one or more orders of magnitude higher. The rationale underlying this stringent EC standard is based on the thesis that “pesticides have no place in drinking water”. However, whilst this is a desirable objective, the reality is that in many countries in Europe, pesticide contamination of raw drinking water sources has occurred above 0.1 µg/l. This in turn has triggered huge investment in treatment plants by water utilities, to try and achieve the pesticide standard in the water supplied to customers.

      Treatment is an extremely expensive option, and even well designed and operated plants cannot necessarily guarantee full compliance with the standard on an absolute basis. More importantly sole reliance on treatment does not reflect the “polluter pays” concept, which recognises that the prime objective in such situations is to control the problem at source by catchment protection, rather than attempt to clean up degraded water resources.

      Nevertheless, whilst meeting pesticide standards by means of source protection must be the prime goal, this is likely to take some time to achieve, and realistically may not be fully possible at all in some situations. Treatment must therefore continue to play a part in meeting pesticide standards, if only as an interim situation until catchment protection measures take effect.

      Pesticide problems should therefore be tackled as part of an integrated strategy, which seeks the fastest, most practicable, and cheapest solution that can be sustained into the future.

Background to the problem of herbicides in water

Severn Trent Water provides drinking water to over seven million people in the UK Midlands, from a variety of water sources. The company has regularly monitored for pesticides in water for a number of years and contamination in excess of the EC standard has been regularly detected in our river derived sources. The failures are, typically, seasonal in nature and tend to reflect peak pesticide usage patterns and rainfall events. Over the last few years, up to 10% of sourcework samples have exceeded 0.1 µg/l, the greatest majority of failures being associated with our lowland river derived abstractions. Maximum pesticide results up to 1-2 µg/l have been found, but typically results are much lower. The most commonly detected pesticides are the non-agricultural herbicides, atrazine, simazine and diuron; and the cereal herbicides mecoprop, MCPA, isoproturon and chlorotoluron.

      By contrast, our other two major resource types, the upland reservoirs and sandstone aquifers, so far appear to be generally much less affected by pesticide contamination. This does not however mean that localised problems have not occurred, or that potentially significant contamination risks do not exist. For example, three of our 170 borehole sources have been significantly affected by herbicide contamination, two of which were believed to be influenced by surface infiltration from a nearby railway, and one which was known to be in hydraulic continuity with a river source which was itself contaminated. This seems to reflect experience elsewhere, that groundwater sources which are either shallow or have some form of fast recharge mechanism are much more vulnerable to pollution than deeper boreholes with thicker unsaturated zones—the more common situation in our region.

      Similarly our upland catchments by comparison with river catchments are relatively small in area, and generally not subject to intensive development or agricultural use. The risks of contamination are however real, since one pollution incident could cause the whole reservoir to breach the pesticide standard. For example, transient pollution was observed in one of our smaller reservoirs, due to a one-off road spraying exercise by the local municipality on some minor roads around the reservoir catchment.

Pesticide use surveys

Before any integrated and sustainable long term strategy for pesticides is developed, it is essential to have a good knowledge of pesticide use within the various catchment areas, and to be able to forecast any medium and long term changes in use pattern. This helps to focus expensive sampling effort on those pesticides most likely to be present and allows more effective targeting of catchment control measures.

      To achieve this awareness of pesticide use we have let a contract with a specialist company which provides us with a regular quarterly listing of estimated product use in each parish (municipal sub-district) in our main catchment areas. This information is derived from the Ministry of Agriculture (MAFF) information on crops grown in each parish and is then related to farming practice information for each crop, gained from approximately 2,000 farms in the UK. Pesticides use is extrapolated from these databases and cross matched to pesticide sales for the area in order to check accuracy.

      While this information is very valuable, it only covers pesticides used for agricultural purposes. Since non-agricultural use also constitutes a significant source of contamination, information on these products are also necessary.

      The prime source of information in the UK is a survey of herbicide use in 1989, carried out by the Government in 1991(1). This showed that, of a total of 550 tonnes used, by far the biggest uses of non-agricultural herbicides were:atrazine 25%, simazine 14%, diuron     12%, 2,4-D 9%, mecoprop 3%, amitrole 7%. Since that time we have in  our area supplemented this information with surveys designed to look at trends in pesticide use, particularly by municipalities which are responsible for weed control on roads, and railway authorities.

      In addition, we assess market, regulatory and environmental pressures which might change the future pattern of product development and use, through contacts with national bodies and trade associations.

      Although accurate figures are not yet available some changes in use are occurring.  For example in the non-agricultural sector we believe that there has been a clear decrease in the use of triazines over the last two years, balanced to some extent by an increase in diuron use. Glyphosate has shown a marked increase in use as organisations try to adopt more environmentally-friendly approaches to weed control.

Pesticide analysis requires specialist staff and complex equipment. Photo: Severn Trent Water

Limits and cost of analysis

UK Regulations require water utilities to develop strategies for monitoring pesticides in water, based on local use patterns and the risk that any particular pesticide product might be present in the water sources.

          It is well known, however, that achieving reliable and accurate analysis of water for the wide range of pesticides currently authorised for use is extremely demanding. For example, for UK regulatory purposes, laboratories must be able analyse pesticides at a level of 1/10 of the standard, i.e. 0.01 µg/l, with a maximum total tolerable error of -/+ 20%.

      Significant advances in analytical technology for pesticides in water have been made in recent years, and our laboratory now routinely analyses for 74 pesticides, as set out in Table 1. Nevertheless there are still a number of pesticides in use for which reliable methods are not available, although further development of new analytical procedures is continually taking place.

      One of the particular difficulties of low level pesticide analysis in water samples is achieving confidence that the detected peak on the separation column is uniquely and accurately ascribed as a particular pesticide. It is not uncommon for example for false positive results to be generated because of co-elution of other substances, or other artefacts of the analytical procedure. For this reason our laboratory now routinely uses mass-spectrometric confirmation  techniques for all suspected positive pesticide detections. Adoption of this approach, whilst expensive and complex, has resulted in a reduction in the number of isolated and spurious detections of probable false results.

      It is important to emphasise that, because of the complexity of modern pesticide analysis, it is a very expensive exercise. During the last year alone, Severn Trent Water spent £1 million on pesticide analysis. This is a very onerous burden for any utility and reinforces the need to target the monitoring effort very precisely to ensure that maximum benefit and information is generated for this investment.

Catchment protection via the Spraysafe campaign

Within our area, one of the major sources of pesticide contamination was known to be from the use of residual herbicides to control weed growth on hard surfaces, particularly roads and railways. We therefore initiated a major campaign entitled “Spraysafe”, to persuade all users to significantly modify weed control practice, in order to reduce or even eliminate pesticide leaching into catchments. The Spraysafe campaign was carried out with the strong help and support of the agrochemical industry, and was generally well received by herbicide users as being positive and constructive.

      The campaign involved a number of overlapping but co-ordinated actions over an initial period of 18 months. It is however intended to maintain and develop the campaign as a permanent feature, as well as extending it to agricultural users. Key points of the campaign included:

1. User surveys/mail shots:  Specific herbicide use and attitude  surveys among municipalities were carried out. This had the parallel advantage of building up a database of over 1,500 individual herbicide users in 205 municipalities, as well as other bodies to whom we could send targeted mailshots and information.

2. Spraysafe charter: Widespread dissemination of a document  highlighting the risks to water catchments of herbicide use, and providing a simple eight-point check list of ways to avoid or minimise water contamination.

3. Spraysafe conferences: Herbicide users, contractors, research agencies and agrochemical companies were invited to a free major conference and exhibition. Two conferences were held and both were significantly oversubscribed. The proceedings included both water industry and agrochemical experts describing the scale of the pesticide problem and practical ways to avoid such pollution.

4. Press campaigns: Alongside the conferences a variety of press and TV coverage was achieved, highlighting the problem and encouraging users that had not already adopted good practice, to follow the example of those that had.

5. Promotion of detailed advice services: One of the key findings of our surveys was that many herbicide users were willing to change weed control practice to avoid water pollution, but had limited access to expert practical advice on how this might be achieved. We therefore worked with a number of major agrochemical companies to encourage the development of training packages using video and other techniques, that were provided free to users. Concurrently the government produced simple codes of good practice on herbicide use(2,3). At the same time a specialist independent advice agency developed a commercial consultancy service.

6. Railway track spraying: Railways need to be kept clear of weeds for safety and other reasons. In the UK this has traditionally been achieved by an annual spraying of all track lanes with a residual herbicide. However given that after spraying, track drainage is likely to be heavily contaminated with herbicide, this could pose a significant threat to adjacent or downstream water abstractions. In consequence following discussion with the national railway company, and their contractors, they have now agreed to spray only non-residual herbicide in areas close to boreholes or river abstractions which are particularly vulnerable to water contamination. 

Overall, this campaign has had a major benefit in raising awareness of the need to significantly change weed control policy in order to reduce water source contamination. Encouraging feedback is being received about changes in practice and some clear reduction in triazine levels is being seen.  By contrast there is some evidence of the growing problem with diuron and contamination by agricultural herbicides remains stubbornly high.

Land use controls in water sensitive areas

Land use activities in catchment areas, especially those close to the water treatment works abstraction point, can significantly affect raw water quality both in terms of pesticide and other contamination. This may be through diffuse pollution (eg agrochemical or farm slurry application to fields, cattle grazing) or point pollution (eg spillage or disposal of chemicals, or fire damage to chemical stores).

      In the case of upland catchment areas, land use activities around impounding reservoirs are likely to have the greatest influence upon water quality. In addition to agricultural pollution, water quality problems arise from soil erosion, deforestation and tourism.

      For each of its impounding reservoirs, Severn Trent is therefore preparing a Catchment Protection Policy Document. This document assesses potential pollution risks from all sources and provides pragmatic advice to minimise these risks. Areas addressed in the document include:

• definition of catchments and identification of landowners;

• control of risks from: animal husbandry; organic fertilisers; chemical fertilisers; pesticides; forestry management; recreation tourism.

For some reservoirs where land is owned by the water company, land use practices can be directly controlled through the farm tenancy agreement. In other areas catchment protection will generally be on a voluntary basis, influenced by personal contact and discussion and supported by information and guidance literature. It is recognised that catchment protection advice is more likely to be accepted if it is practical, can save the land user money or protect him from risk of prosecution from pollution control agencies. By enrolling the assistance of agricultural and land management experts in the formulation of policy, a realistic and sensible approach to water quality protection is assured.

The impact of regulatory controls

In the UK the prime regulatory control over pesticides is exercised through an official use approval system. Only pesticide active ingredients and formulations specifically approved may be used, and only then for specified applications.

      The two pesticides most commonly failing drinking water  standards over the last few years in England and Wales were the triazine herbicides, atrazine and simazine. Unlike many countries the overwhelming use of these substances in the UK was for non-agricultural purposes. Because of their widespread detection in drinking water, and because alternative products were available, following pressure from the water industry and other organisations the UK Government in 1992 therefore announced the withdrawal of use authorisation for atrazine and simazine for non-agricultural purposes from 1993.

      This announcement reinforced the message we were promoting via the Spraysafe campaign, and provided a good example of how both regulatory and voluntary approaches could work together. The implicit threat that further use restriction might be introduced if water contamination was not reduced provided a powerful incentive to agrochemical producers and users to work with us in promoting voluntary changes in practice.

      Nevertheless we found it vital that the Spraysafe campaign disseminated a clear message about pesticide use. For example, we regularly came across the view that it was only triazine herbicides that were causing a problem, and therefore ill-informed users simply planned to switch to an equally mobile and polluting herbicide such as diuron, which was already beginning to appear in drinking water sources with the same regularity as triazines. Our campaign message was therefore that all pesticide use potentially causes water contamination risks, and hence a much more careful and selective approach to all weed control was necessary.

      The withdrawal of triazines for non-agricultural purposes did not in practice have a major impact on the ability of users to control weeds since alternative approaches were readily available. It will be interesting to see whether regulatory changes are as easy to promote if there are, for example, major impacts in crop protection control.

Vulnerability to pesticide contamination varies

It is becoming increasingly clear that the susceptibility of water catchments to pesticide contamination varies considerably. Whilst blanket controls and encouragement of general good practice measures will be of major benefit in reducing overall pesticide contamination levels, high degrees of protection will only be possible if this is supplemented by targeted controls in high risk areas. Similarly it is important that pesticide manufacturers are encouraged in the longer term to develop pesticides that are less environmentally mobile, either because they degrade and/or bind tightly to soil; and/or can be used at lower application rates.

Environmental fate: a developing field

All of these approaches require a much better understanding of the environmental fate and behaviour of different pesticides. Much research on this topic is taking place, not least by manufacturers themselves. It is now clear, for example, that soil type can play a critical role in determining the environmental fate of any applied pesticide, and that factors which promote bypass of the soil zone, such as field drainage, soil fissuring, or the bulk disposal of pesticides, can dramatically affect the level of  water pollution.

      For that reason, in conjunction with the Soil Survey and Land Research Centre (SSLRC), we have developed a customised, computer based catchment planning system for our region, known as CATCHIS. This holds data on soil types, climatic characteristics, surface hydrographic boundaries, aquifer recharge areas and other surface features such as rivers, roads and railways, superimposed on a base map of our area(4). From this, by the use of simple models which integrate local climate, soil, aquifer, and pesticide characteristics, we can plot those areas of highest pesticide leaching risk which are closest to our abstraction points. This in turn will allow us to encourage the focusing of particularly stringent pollution control measures in those areas, for example by seeking the total avoidance, if possible, of those pesticides known to be very mobile and persistent.

      In the future it is expected that this system will be integrated with a national relational database system known as SEISMIC being developed by SSLRC, the British Agrochemical Association and the Pesticides Safety Division of MAFF(5). SEISMIC will provide the spatial soil, cropping and agro-climatic data necessary to identify vulnerable areas in relation to specific crops, and then provide the appropriate soil and weather data necessary to undertake the detailed modelling and prediction of likely pesticide level in soil leachate or runoff. It will thus enhance the Severn Trent catchment planning system (CATCHIS) by refining both the identification of vulnerable areas and the prediction of pollution risk within those areas.

Water treatment: residual removal

It will take some time to secure consistently robust catchment protection which is sufficient to maintain raw waters in compliance with the EC standard. Realistically this objective may never be achieved in some catchments; particularly in river catchments. In such circumstances the water utility will need to install water treatment plant to ensure compliance with the standard in the water delivered to the customer.

      Much work has been done on pesticide removal treatment; the techniques most commonly used in many water works being based on granular activated carbon (GAC), and/or ozone, with or without peroxide addition. These processes, of course, can also provide other water quality benefits but the scale, size and cost of plant required is often dictated by pesticide contamination levels.

      The choice of which process or combination of processes to use will depend very much on local circumstances. In our area for example, since pesticide contamination levels were variable and not consistently at high concentrations, our prime treatment strategy on surface sources is to rely on secondary GAC adsorbers with an empty bed contact time between 10 and 20 minutes depending on the plant. This generally has been shown to be successful, although it may potentially require relatively frequent GAC regeneration (a year or less) and will not necessarily guarantee compliance if challenged by a very high raw water pesticide  peak. Nevertheless, if our Spraysafe campaign achieves the success that is hoped, then GAC is likely to be the cost effective option for treatment of residual pesticides in raw water at many plants. One of the advantages of GAC for example is that it has a degree of flexibility in terms of regeneration frequency that allows reduction in raw water pesticide levels to be matched by reduced GAC operating costs.

      Ozone, with or without peroxide, may however also be cost effective for pesticide removal in some situations, particularly where pesticide levels in raw water are higher. We already have one operational ozone plant and another under construction. We also have a major advanced water treatment pilot plant, to allow different combinations of pesticide removal techniques to be optimised.

Conclusions

The problem of pesticide contamination of drinking water is complex, dynamic and expensive to resolve. Solutions must therefore address issues over both short and longer timescales and involve co-operation between all interested parties. It is our experience that a constructive approach can be developed, but this requires sustained effort on all sides with the water utility tending to act as the prime focus for co-ordination. Nevertheless success is unlikely to be achieved by voluntary action alone, and therefore a clear framework of statutory control needs to be established within which the needs of both crop protection and water protection can be recognised. In the longer term it is reasonable to assume that a more sophisticated and carefully controlled approach to pesticide use could dramatically reduce the incidence of raw water contamination by pesticides. However, while this should be the prime objective, there may be circumstances when this is not technically or economically practicable, in which case some modest degree of water treatment may still be required.  

• The authors are grateful for the constructive support from and discussion with colleagues in the water and agrochemical industries, which underpinned much of the work described in this paper. Particular thanks go to staff at Severn Trent Water for their professional input to development and implementation of the pesticide strategy, including the skilled analysts at Severn Trent Laboratories; Andrew Heather, Chris King-Turner and Paul Williams for their support in the Spraysafe campaign; and John Hollis and his team at SSLRC for development of the CATCHIS system.

• The authors acknowledge permission from Severn Trent Water to publish this paper, but any views expressed are personal and not necessarily those of the company.    

• Dr. Bob Breach is Quality Planning Manager and Mike Porter Quality Planner (Water Supply) at Severn Trent Water Ltd,  2297 Coventry Rd, Birmingham, B26 3PU, UK.

References:

1. The use of herbicides in non-agricultural situations in England and Wales, prepared for the UK Department of Environment by the Foundation for Water Research, 1991.

2. Weed Control and Environmental Protection, UK Department of the Environment, 1992.

3. Guidance for Control of Weeds on Non-Agricultural Land, UK Department of Environment, 1992.

4. Hollis, J.M., Mapping the Vulnerability of Aquifers and Surface Waters to Pesticide Contamination at the National/Regional Scale, BCPC Monograph No. 47—Pesticides in Soils and Water, 1991, pp 165-174.

5. Hollis, J.M., The Development of Integrated Database Systems for Modelling Pesticide Environmental Fate and Behaviour, BCPC Conference—Weeds, Proceedings, 1993.

[This article first appeared in Pesticides News No. 22, December 1993, pages 6-9]